ψ %HOJLFD3V\FKRORJLFD

Heider, N., et al. (2018). Body Dissatisfaction Revisited: On the

Importance of Implicit Beliefs about Actual and Ideal Body Image
PsychologicaBelgica,58(1),pp. 158–173,DOI:https://doi.org/10.5334/
pb.362

RESEARCH ARTICLE

Body Dissatisfaction Revisited: On the

Importance of Implicit Beliefs about Actual
and Ideal Body Image
Niclas Heider, Adriaan Spruyt and Jan De Houwer

Body dissatisfaction (i.e., a negative attitude towards one’s own physical a

­ ppearance)
is assumed to originate from a perceived discrepancy between the actual physical
appearance (i.e., actual body image) and the desired ideal state of the body (i.e.,
ideal body image). We assessed implicit beliefs about these two aspects of the
body image independently using two Relational Responding Tasks (RRT) in a sample
of participants who were either low or high in explicitly reported body dissatis-
faction. As hypothesized, differences in body dissatisfaction exerted a differential
influence on the two RRT scores. The implicit belief that one is thin was less
pronounced in participants who were strongly dissatisfied with their body ­relative
to participants who were more satisfied with their body. The implicit desire to
be thin (i.e., thin ideal body image), in contrast, tended to be more pronounced
in participants who exhibited a high degree of body dissatisfaction as compared
to participants who exhibited a low degree of body dissatisfaction. Hierarchical
regression analyses also revealed that the RRT scores were predictive of self-
reported body dissatisfaction, even over and above the predictive validity of some
(but not all) explicit predictors of body dissatisfaction that were included in the
present study. More generally, these findings contribute to the empirical validation
of the RRT as a measure of implicit beliefs in the context of body dissatisfaction.

Keywords: Implicit measures; relational responding task; body dissatisfaction;

eating disorder

Body (image) dissatisfaction can be defined
as the negative attitude towards one’s own
body resulting from a perceived discrep-
ancy between the actual body image (i.e.,
Department of Experimental Clinical and Health
Psychology, Ghent University, BE
Corresponding author: Adriaan Spruyt
(Adriaan.Spruyt@UGent.be)
perceptions, thoughts, and feelings con-
cerning one’s actual physical appearance;
e.g., Cash, 1990) and the ideal body image
(i.e., internalized ideals about one’s physi-
cal appearance; e.g., Cooper & Taylor, 1988;
Strauman, Vookles, Berenstein, Chaiken, &
Higgins, 1991; Williamson, Gleaves, Watkins, &
Schlundt, 1993). Body dissatisfaction has
been identified as one of the key factors for
Heider et al: Body Dissatisfaction Revisited
dieting behavior, negative affect, and the
causation and maintenance of eating disor-
ders (American Psychiatric Association, 2013;
Fairburn & Harrison, 2003; Stice, 2001, 2002).
Accordingly, over last few decades, behavioral
scientists have invested a great deal of effort
in the development of measures of body dis-
satisfaction (Allebeck, Hallberg, & Espmark,
1976; Bessenoff & Sherman, 2000; Bluemke &
Friese, 2012; Degner & Wentura, 2009;
Freeman, Thomas, Solyom, & Hunter, 1984;
Heider, Spruyt, & De Houwer, 2015; Juarascio
et al., 2011; Parling, Cernvall, Stewart, Barnes-
Holmes, & Ghaderi, 2012; Roddy, Stewart, &
Barnes-Holmes, 2010, 2011; Slade & Russell,
1973). Most of these instruments, however,
are self-report measures (i.e., questionnaires)
and are thus limited in two ways. First, it is
well-known that self-report measures are
susceptible to effects of social desirability
and impression management (Cronbach,
1990; Holtgraves, 2004). For example, when
completing a body-dissatisfaction question-
naire, anorectic patients may be motivated
to respond untruthfully because of far-
reaching therapeutic consequences (e.g.,
compulsory admission). Second, as discussed
by Greenwald and Banaji (1995), self-report
measures are, by definition, unsuited to
capture traces of past experiences that
are introspectively unidentified (e.g., the
­accidental but repeated exposure to the thin
ideal on a social media).
To resolve both problems, behavioral
scientists have started to develop diagnostic
instruments that allow for an assessment
of psychological constructs (e.g., attitudes,
biases, beliefs) without relying on self-reports.
More specifically, these ‘implicit measures’
aim at capturing inter-individual differences
under automaticity conditions, for example
by capitalizing on the well-known principle
of automatic response facilitation/interfer-
ence, the presentation of construct-relevant
stimuli under subjective recognition thresh-
olds, and/or the implementation of strict
response deadlines (for an extensive discus-
sion of the meaning of the concept ‘implicit
159
measure’, see De Houwer, Teige-Mocigemba,
Spruyt, & Moors, 2009).
Roughly, the class of implicit measures
can be divided into two broad subclasses.
Associative implicit measures, like the
(standard) evaluative priming task (EPT;
Fazio, Jackson, Dunton, & Williams, 1995),
the Implicit Association Test (IAT; Greenwald,
McGhee, & Schwartz, 1998), or the Affect
Misattribution Procedure (AMP; Payne,
Cheng, Govorun, & Stewart, 2005), are
designed to assess the associative strength
between concepts in memory (Hughes,
Barnes-Holmes, & Vahey, 2012). Associative
implicit measures can, for instance, be used
to capture the extent to which a certain class
of stimuli (e.g., spiders) is associated with a
positive or negative valence.
In many cases, however, it is not only impor-
tant to examine to what extent two concepts
are related in memory but also the precise
way in which they are related (e.g., Heider
et al., 2015; Remue, De Houwer, Barnes-
Holmes, Vanderhasselt, & De Raedt, 2013;
Remue, Hughes, De Houwer, & De Raedt,
2014). In these cases, the use of relational
implicit measures, like the Implicit Relational
Assessment Procedure (IRAP; Barnes-Holmes
et al., 2006) and the Relational Responding
Task (RRT; De Houwer, Heider, Spruyt, Roets, &
Hughes, 2015) is more appropriate. These
measures were designed specifically to allow
for an assessment of the way in which indi-
viduals tend to relate two stimuli to one
another automatically. In the context of
spider fear, for example, respondents who
complete an RRT may be presented with
relational information concerning spiders
(e.g., the statements ‘I like spiders’ and ‘I hate
spiders’). Their task is to respond ‘true’ or
‘false’ on the basis of an instructed response
rule. For example, participants may be asked
to respond as if they despise spiders in one
block of trials (e.g., respond ‘true’ to the
statement ‘I hate spiders’) and to respond
as if they like spiders in a second block
of trials (e.g., respond ‘false’ to the state-
ment ‘I hate spiders’). By examining which
160
relational information (e.g., ‘like’, ‘hate’)
results in optimal task performance given a
specific response rule, one can learn about
the precise way in which ­ respondents
tend to relate specific stimuli (e.g., ‘I’,
‘spiders’). Moreover, given that participants
are required to respond as fast as possible
while acting as if they endorse a certain
belief, it can be argued that RRT scores
are more implicit in terms of speed and
unintentionality than traditional explicit
­ on the definition of body dissatisfaction
measures such as questionnaires (for an
extensive discussion of the ‘implicit’ nature
of the RRT, see Heider et al., 2015).
As argued above, body dissatisfaction
depends on beliefs about the actual and the
ideal body image. Crucially, these two types
of body image differ only in how the concepts
‘self’ and ‘body-size’ are related to another.
Specifically, beliefs about the actual state
of one’s body involve a descriptive relation
(e.g., ‘I am thin’) whereas beliefs about the
ideal state of one’s body involve a relation of
desirability (e.g., ‘I want to be thin’). Therefore,
relational implicit measures should be best
suited to assess these different beliefs at the
implicit level.
Preliminary evidence for this assertion was
recently published by Heider et al. (2015).
In their study, participants completed two
IRAPs, one to capture actual body image
(hereafter referred to as actual-IRAP) and
one to capture ideal body image (hereafter
referred to as ideal-IRAP). In both tasks, two
stimuli were presented on each trial. The
so-called sample stimuli (i.e., ‘I am’ vs. ‘I am
not’) were presented at the top of the com-
puter screen. The so-called target stimuli (i.e.,
words referring to thinness and overweight)
were presented in the center of the com-
puter screen. Crucially, the combinations
of the two stimuli were either congruent or
incongruent with being thin (e.g., ‘I am’ +
‘skinny’ and ‘I am’ + ‘chubby’, respectively). In
one set of trials, participants were asked to
respond as if they believed themselves to be
thin by selecting the appropriate response
option (i.e., ‘true’ or ‘false). In as second set
Heider et al: Body Dissatisfaction Revisited
of trials, they were asked to respond as if they
believed themselves to be overweight. The
ideal-IRAP was identical to the actual-IRAP
except for the fact that participants were (a)
presented with stimulus combinations that
reflected the desire to be thin or overweight
(e.g., ‘I want to be + skinny’ and ‘I want to
be + chubby’, respectively) and (b) required
to respond as if they desired to be thin or
overweight in different sets of trials. Based
as the negative attitude towards one’s own
body resulting from the perceived discrep-
ancy between actual and ideal body image,
Heider et al. (2015) expected the scores of
actual-IRAP and ideal-IRAP to vary as a func-
tion of self-reported body dissatisfaction. As
anticipated, results indicated that the belief
to be thin (as measured by the actual-IRAP)
was more pronounced in participants who
were low in body dissatisfaction as compared
to participants who were high in body dis-
satisfaction. Scores of the ideal-IRAP, in con-
trast, revealed that the desire to be thin was
less pronounced in participants low in body
dissatisfaction as compared to participants
high in body dissatisfaction. These observa-
tions support the validity of the IRAP as a
measure of specific body-related beliefs.
In the present report, we revisit the implicit
measurement of ideal and actual body image,
for two reasons. First, the IRAP is a challenging
task to complete, even in respondents who
are highly familiar with computerized (reac-
tion-time) tasks. The first phase of an IRAP
typically consists of a n ­ umber of practice
blocks in which participants are required to
reach a certain threshold in terms of average
speed (typically between 2000 ms and 5000
ms) and accuracy (typically between 65 %
and 80 %, for a review, see Hughes & Barnes-
Holmes, 2013). After the practice phase, par-
ticipants can continue with the actual test
phase of the IRAP only if both criteria are
met. Compared to the normal error rates and
average response speed in the associative
implicit measures, these criteria are rather
liberal. Still, attrition rates of 20% or more
Heider et al: Body Dissatisfaction Revisited
are common in IRAP research (e.g., Remue
et al., 2013; see Hughes & Barnes-Holmes,
2013, Table 1, for an overview). Clearly, if
it is to be of value outside of the laboratory,
implicit measures of ideal and actual body
should preferably be less demanding to
complete than the standard IRAP.
As second point of concern, one may note
that the reliability of the IRAP is limited. In a
review of 27 IRAP studies (Hughes & Barnes-
Holmes, 2013), good internal consistency
estimates were observed only if the training
criteria for the practice blocks were quite
liberal (i.e., mean response latencies between
3000 ms and 5000 ms, error rates between
65 % and 70 %; e.g., Barnes-Holmes,
Murtagh, Barnes-Holmes, & Stewart, 2010;
Campbell, Barnes-Holmes, Barnes-Holmes, &
Stewart, 2011; Drake et al., 2010). This obser-
vation may be problematic because social
desirability and impression management
concerns are more likely to affect task perfor-
mance as participants are given more time to
respond on each individual trial.
Studies using the RRT (e.g., De Houwer
et al., 2015), on the other hand, suggest that
the RRT is user-friendly (i.e., attrition rates
as low as 5%, mean response times well
below 2000 ms, and error rates well below
20 %) whilst its reliability is still acceptable
(e.g., Rsb = 0.64 in De Houwer et al., 2015).
Accordingly, in the present study we re-
addressed the implicit assessment of body
dissatisfaction by using the RRT. To capture
beliefs about actual body image, participants
were presented with statements like ‘I p ­ ossess
a slim body’ and ‘I see myself as a fat person’
(hereafter referred to as the actual-RRT).
To capture beliefs about ideal body image,
participants were presented with state-
ments like ‘I wish I was thinner’ and ‘I strive
to weigh more’ (hereafter referred to as the
actual-RRT). In line with Heider et al. (2015),
we hypothesized that participants would dif-
fer in their RRT scores as a function of their
level of explicit body dissatisfaction. More
specifically, given that our sample consisted
of young female adults who typically desire
161
to be thin rather than overweight (i.e., thin-
ideal internalization; e.g., J. K. Thompson &
Stice, 2001), we expected the belief to be
thin to be more pronounced in participants
low in body dissatisfaction as compared to
participants high in body dissatisfaction.
In contrast, we expected the desire to be
thin to be less pronounced in participants
low in body dissatisfaction as compared to
participants high in body dissatisfaction.
Participants also completed a number of
explicit measures of actual and ideal body
image. First, they were asked to rate their
endorsement of the statements used in the
two RRTs. Second, they were asked to choose
from a range of schematic body images the
images that corresponded with their actual
and ideal body (i.e., the Contour Drawing
Rating Scale, CDRS, M. A. Thompson & Gray,
1995). As participants were healthy univer-
sity students who were tested anonymously,
we expected them to respond truthfully on
the explicit measures. We thus expected
implicit and explicit measures of actual and
ideal body image to correlate.
Method
Ethics Statement
Participants gave written informed consent
prior to their participation and received
course credit (n = 62) or payment of €7
(n = 6) in exchange for their participation.
The experiment was approved by the ethics
committee of the Faculty of Psychology and
Educational Sciences of Ghent University.
Participants
At the beginning of the academic year, 468
students at Ghent University completed the
body dissatisfaction subscale of the Eating
Disorders Inventory (EDI; Garner, Olmstead, &
Polivy, 1983) during an online screening
study that involved multiple question-
naires. To ensure that our sample included
participants who were either low or high in
body dissatisfaction, we invited all female
­students who had scored within the first and
fourth quartile of the total EDI distribution
 162

Table 1: Correlations between measures.

M SD min max 1 2 3 4 5 6 7 8
1 Body Dissatisfaction (EDI) 33.17 14.58 12 54 – 0.54*** −0.36** 0.28* 0.67*** −0.06 −0.85*** 0.80***
2 BMI 21.71 3.17 17.26 31.02 – −0.34** 0.00 0.82*** 0.48*** −0.76*** 0.62***
3 actual-RRT 0.08 0.30 −0.46 0.71 − 0.01 −0.28* −0.10 0.30** −0.15
+
4 ideal-RRT −0.05 0.33 −0.63 0.63 – 0.03 −0.36** −0.22 0.13
5 actual-CDRS 4.98 1.69 2 9 – 0.45*** −0.80*** 0.70***
6 ideal-CDRS 3.61 0.97 2 6 – −0.07 −0.04
7 Statements (Actual-RRT) 2.98 1.27 1 5 – −0.85***
8 Statements (Ideal-RRT) 4 0.83 1.2 5 –

Note. actual = actual body image; ideal = ideal body image; RRT = Relational Responding Task; CDRS = Contour Drawing Rating Scale; EDI = Body
dissatisfaction subscale of the Eating Disorders Inventory; BMI = Body Mass Index.
+
p < .10; *p < .05; **p < .01; ***p < .001.
Heider et al: Body Dissatisfaction Revisited
Heider et al: Body Dissatisfaction Revisited
to participate in an individual lab session
(N = 143). Note, however, that the invita-
tion to participate in the experiment was
sent anonymously via an online recrutement
system. It was thus necessary to administer
the EDI for a second time during the actual
experimental session to identify group
membership. In total, 68 female students
(M = 18.72 years, SD = 2.12) responded to
our invitation and participated. All partici-
pants were Dutch speakers and had normal
or corrected-to-normal vision.
Measures
RRTs
Participants completed two RRTs, one to
­capture implicit beliefs about actual body
image (i.e., actual-RRT) and one to capture
implicit beliefs about ideal body image
(i.e., ideal-RRT). In line with De Houwer
et al. (2015), two types of stimuli were used
in both RRTs, inducer words and target
­statements. The same set of 10 inducer words
was used in both RRTs, five of which were
synonyms of ‘true’ and five of which were
synonyms of ‘not true’ (see Table 2 in the
Appendix). Two sets of 20 sentences were
used as target statements, one set for each
RRT. Target statements for the actual-RRT
related the concepts ‘thin’ and ‘overweight’
to the concept ‘self’ in a descriptive way. Five
statements referred to the belief to be thin
(e.g., ‘I possess a slim body’) and five state-
ments referred to the belief to be overweight
(e.g., ‘I weigh too much’). Negations of these
ten statements led to the creation of 10
additional target statements that referred
to the belief not to be overweight (e.g., ‘I do
not weigh too much’) and the belief not to
be thin (e.g., ‘I do not possess a slim body’),
respectively. Negations were included to
ensure that participants would be required
to process the meaning of the entire state-
ment (and not just a subset of words) in
order to respond correctly. For the ideal-RRT,
the same target concepts were used, but now
the target statements specified a relation of
desirability. Five statements referred to the
desire to be thin (e.g., ‘I desire to weigh less’)
163
and five target statements referred to the
desire to be overweight (e.g., ‘I strive to
weigh more’). Again, negations of these ten
statements were used as 10 additional target
statements that referred to the desire not to
be overweight (e.g., ‘I don’t strive to weigh
more’) and the desire not to be thin (e.g.,
‘I don’t desire to weigh less’; for the com-
plete list of the target statements and their
English translations, see Tables 3 and 4 in the
Appendix).
In both RRTs, on each trial either one of the
inducer words or one of the target statements
was presented on the computer screen. Trials
can therefore be divided into inducer trials
and target trials. On all trials, participants
were instructed to categorize the presented
item as ‘true’ or ‘not true’ by pressing the
right or left ctrl-key of the keyboard, respec-
tively. It may be noted that the inducer trials
were included to ensure that the responses
were endowed with the meaning ‘true’ and
‘false’, similar to other implicit measures
such as the Extrinsic Affective Simon Task
(De Houwer, 2003).
Each RRT comprised seven blocks. In line
with the typical IAT approach (Greenwald et
al., 1998), three blocks were required to famil-
iarize participants with the different response
tasks (i.e., Blocks 1, 2, and 5). The r­ emaining
blocks (i.e., Blocks 3, 4, 6, and 7, sometimes
also referred to as ‘mixed blocks’) were
used for the actual assessments. In Block 1
(40 trials), each of the ten inducer words
was presented four times. Participants were
asked to classify these words as ‘true’ of ‘false’
as fast as possible. In Block 2 (40 trials), the
20 target statements were presented twice.
Participants were asked to respond as if they
were thin (in the actual-RRT) or wanted to be
thin (in the ideal-RRT). For example, in the
actual-RRT, the response ‘false’ was required
when the statement ‘I weigh too much’ was
presented. In blocks 3 and 4 (40 trials each),
the ten inducer words were presented twice
and the 20 target statements were presented
once. Participants were asked to respond in
accordance with the response rules prac-
ticed during the preceding blocks. Block 5
164
(40 trials) was identical to Block 2, except for
a reversal of the response rule. Participants
were now asked to respond as if they were
overweight (in the actual-RRT) or wanted to
be overweight (in the ideal-RRT). For example,
in the ideal-RRT, the response ‘true’ was
required when the statement ‘I strive to
weigh more’ was presented. Blocks 6 and 7
(40 trials each) were identical to Block 3 and 4,
but participants were asked to respond in
line with the response rule practiced in
Block 5. The order of the different items was
random within each block, with the restric-
tion that the same item could not be repeated
on consecutive trials.
Each trial started with the p ­resentation
of an item (i.e., inducer word or target
­statement) in the middle of the computer
screen in Tahoma letter font, 28-point font
size. Inducer words were presented in white
in both RRTs, whereas the target statements
were presented in orange (actual-RRT) or blue
color (ideal-RRT). Different colors were used
to increase awareness of the fact that differ-
ent target statements were presented in the
two RRTs. Items remained on screen until the
correct response was registered. Incorrect
responses were signaled by the presentation
of a red X (Arial, 72-point font size) below
the item until participants gave the correct
response. The next trial started 750 ms after
registration of the correct response. The RRTs
were presented on a 17-inch LCD screen (60
Hz, 1440 × 900 pixels) and were written in
Affect 4.0 (Spruyt, Clarysse, Vansteenwegen,
Baeyens, & Hermans, 2010).
Self-report measures
Participants rated each of the 40 target
statements used in the RRTs on a 5-point
scale, ranging from 1 (disagree completely)
to 5 (agree completely). For each individual
participant and each class of target state-
ments, these ratings were aggregated to
obtain an explicit measure of actual body
image and ideal body image. Actual and ideal
body image were also measured using the
female version of the Contour Drawing Rating
Scale (CDRS; M. A. Thompson & Gray, 1995).
Heider et al: Body Dissatisfaction Revisited
The CDRS consists of nine schematic (female)
figures of varying sizes ranging from under-
weight (1) to overweight (9). Participants
completed the CDRS twice, once to indicate
their actual body image and once to indi-
cate their ideal body image. Explicit body
dissatisfaction was assessed by means of the
body dissatisfaction subscale of the Eating
Disorder Inventory (EDI, 9 items; Garner
et al., 1983), which has excellent ­psychometric
qualities (e.g., Clausen, Rosenvinge, Friborg, &
Rokkedal, 2011). Finally, we computed the
Body Mass Index (BMI) for each participant
using self-reported weight and height. We
did not collect objective measures of weight
and height as we observed an almost perfect
correlation between factual and self-reported
BMI measures in prior research (i.e., r = 96,
Heider et al., 2015).
Procedure and Group Assignment
All participants were tested individually
during an experimental session that on
­average lasted 35 minutes. Because the focus
of the present research was on the ­usefulness
of the actual-RRT and the ideal-RRT, we
wanted to ensure that performance in these
tasks was not influenced by the prior com-
pletion of explicit measures. Accordingly, the
experimental session always started with the
completion of the two RRTs (in a counter-
balanced order). Subsequently, participants
were asked to rate the target statements used
in the two RRTs and to complete the EDI
and CDRSs for actual and ideal body image.
Finally, participants were asked to report
their weight and height for the calculation
of the BMI.
As sampling was conducted anonymously
via an online recruitment system, group
assignment was based on the EDI ratings
collected during the actual lab session.
Participants were assigned to either the low
or the high body dissatisfaction group by
means of a cluster analysis. In the low body
dissatisfaction group (n = 31), the mean EDI
score was 19.1 (SD = 4.6, min = 12, max = 32).
In the high body dissatisfaction group
(n = 33), the mean EDI score was 46.4
Heider et al: Body Dissatisfaction Revisited
(SD = 5.2, min = 36, max = 54). Both groups
differed significantly in terms of their mean
EDI score, t(62) = 22.35, p < .001. There was
no overlap between groups in terms of the
EDI score. In comparison to the EDI data
collected by Clausen et al. (2011), the mean
level of body dissatisfaction observed in the
low body dissatisfaction group was already
slightly elevated relative to normal controls
(i.e., 19.1 vs. 15.3). In the low body dissatis-
faction group, the mean level of body dissat-
isfaction was very high, even compared to a
sample of eating disorder patients (i.e., 46.4
vs. 27.9).
Results
Data Preparation
The data of two participants were excluded
from the analyses because their mean reaction
times in both tasks (2428 ms and 2629 ms,
for the actual-RRT; 2862 ms and 2838 ms, for
the ideal-RRT) exceeded our cutoff criterion
of 2.5 standard deviations above the grand
mean of the respective tasks (actual-RRT: M =
1292 ms, SD = 369 ms; threshold = 2215 ms;
ideal-RRT: M = 1521 ms, SD = 417 ms; thresh-
old = 2563 ms; see Ratcliff, 1993). We also
excluded the data of two other participants
whose error rates in one of the two tasks
(i.e., 27.5 % and 35.0 %) exceeded the ­cutoff
criterion of 2.5 standard deviations above
the grand mean of that task (actual-RRT: M =
11.2 %, SD = 5.3 %; threshold = 24.4 %; ideal-
RRT: M = 11.3 %, SD = 6.7 %; threshold =
28 %; see Ratcliff, 1993). The mean reaction
time on target trials in the actual-RRT was
1284 ms (SD = 341 ms), with participants
responding incorrectly on 10.9 % (SD =
4.8 %) of the target trials. In the ideal-RRT,
participants needed, on average, 1517 ms
(SD = 394 ms) to respond on target trials and
the error rate was 10.8 % (SD = 6 %).
For each participant and each version of
the RRT, the raw response latencies obtained
in the diagnostic blocks (i.e., Blocks 3, 4, 6,
and 7) were transformed into D scores using
the D1 algorithm described by Greenwald,
Nosek, and Banaji (2003). The D1 a­ lgorithm
was chosen because reaction times were
165
recorded until a correct response was
detected, thus removing the need for the
use of a penalty for incorrect responses.
Following the guidelines of Greenwald
et al. (2003), the calculation of the D1 scores
involved two steps. First, both for the actual-
RRT and the ideal-RRT, separate D1 scores
were calculated for the first and the second
half of the diagnostic blocks (i.e., Blocks 3/6 vs.
Blocks 4/7). The mean response latencies
observed in Blocks 3 and 4 were thus sub-
tracted from the mean response latencies
observed in Blocks 6 and 7, respectively, and
each difference score was divided by the
standard deviation of the respective response
latencies. In a second step, for each RRT, an
overall D1 score was computed by averaging
the two D1 scores obtained for each pair
of blocks. In sum, for the actual-RRT, the
D1 scores were computed such that higher
scores were indicative of the (implicit) belief
to be thin. Similarly, for the ideal-RRT, the
D1 scores were computed such that higher
scores were indicative of a more pronounced
(implicit) desire to be thin.
For the actual-RRT, D1 scores ranged from
–0.46 to 0.71, with a mean score of M = 0.08
(SD = 0.30), which differed from zero, t(63) =
2.07, p = .043, d = 0.26. For the ideal-RRT,
D1 scores ranged from –0.63 to 0.63, with a
mean score of M = –0.05 (SD = 0.33), which
did not differ from zero, t(63) = 1.33, p =
.189, d = 0.17.
Effects at the Group Level
RRTs
To investigate whether the RRT scores were
dependent upon the degree of self-reported
body dissatisfaction, the overall RRT scores
were submitted to a 2 (RRT: actual vs. ideal)
× 2 (body dissatisfaction: high vs. low) ANOVA.
As expected, we found a significant interac-
tion of body dissatisfaction and RRT, F(1, 62) =
8.80, p = .004, η p2 = .12. Participants low
in body dissatisfaction scored higher on the
actual-RRT than participants high in body
dissatisfaction, 0.18 vs. –0.02, t(62) = 2.72,
p = .008, d = 0.68. Conversely, in absolute terms,
participants high in body dissatisfaction
166
scored higher on the ideal-RRT than partici-
pants low in body dissatisfaction, 0.00 vs.
–0.11, t(62) = 1.40, p = .166, d = 0.35. In
addition, we found a significant main effect
2
of RRT, F(1, 62) = 6.90, p = .011, η p = .10,
indicating higher scores on the actual-RRT as
compared to the ideal-RRT, 0.08 vs. –.05. No
other effects were significant, all Fs < 1, all
ps > .475.
To examine the reliability of the RRTs, reli-
ability coefficients were estimated using a
bootstrap procedure. For each RRT and each
of 100 random-splits of the data, the correla-
tion across participants between the two RRT
scores was calculated. Correlations were then
averaged. This procedure resulted in spear-
man-brown corrected mean split-half corre-
lations of Rsb = 0.49 and Rsb = 0.57, for the
actual-RRT and the ideal-RRT, respectively.
Explicit ratings of the target statements of the
RRTs
To investigate whether the explicit ratings
of the target statements used in the RRTs
were also dependent upon the degree of self-
reported body dissatisfaction, they were used
as dependent variables in a 2 (type of target
statement: actual body image vs. ideal body
image) × 2 (body dissatisfaction: high vs. low)
ANOVA. Results mirrored those of the RRT
scores. We found a significant interaction
of type of target statement and body dis-
satisfaction, F(1, 62) = 135.09, p < .001,
η p2 = .69, indicating that participants high
and low in body dissatisfaction differed
in their ratings of the two types of target
statements. Specifically, the explicit endorse-
ment of the belief to be thin was more
pronounced in participants who were low
in body dissatisfaction as compared to
participants who were high in body dissatisfac-
tion, 4.09 vs. 1.93, respectively, t(62) = 12.97,
p < .001, d = 3.24. In contrast, ratings
revealed that the explicit desire to be thin
was more pronounced in participants who
were high in body dissatisfaction as com-
pared to p ­ articipants who were low in body
dissatisfaction, 4.55 vs. 3.42, respectively,
­ between the RRT scores, the CDRS scores of
t(62) = 7.46, p < .001, d = 1.86. Mimicking
Heider et al: Body Dissatisfaction Revisited
the RRT data, we also found a significant
main effect of the type of statement,
F(1, 62) = 47.83, p < .001, η p2 = .44.
Overall, the explicit belief to be thin was
more ­pronounced than the explicit desire
to be thin, 4.00 vs. 2.97. Finally, we found a
­significant main effect of group, F(1, 62) =
49.30, p < .001, η p2 = .44, indicating that
participants who were high in body dissat-
isfaction were on average more extreme in
their rating of the target statements than
participants who were low body in body
­
dissatisfaction, 3.24 vs. 3.75.
CDRS scores
A 2 (type of CDRS: actual body image vs. ideal
body image) x 2 (body dissatisfaction: high vs.
low) ANOVA revealed a significant ­interaction
between the type of CDRS and body dissatis-
faction, F(1, 62) = 66.45, p < .001, η p2 = .52.
In line with the results reported above, the
mean score on the actual-CDRS was higher
in participants who reported a high degree
of body dissatisfaction as compared to par-
ticipants who reported a low degree of body
dissatisfaction, 6.03 vs. 3.87, respectively,
t(62) = 6.59, p < .001, d = 1.65. In c­ ontrast,
there was no evidence of a difference
between the two groups in terms of the ideal-
CDRS (i.e., the mean score was 3.61 in both
groups). We also found a significant main
effect of the type of CDRS, F(1, 62) = 6.90,
2
p < .05, η p = .10. Overall, the scores on the
actual-CDRS were higher than the scores on
the ideal-CDRS, 4.98 vs. 6.31. The main effect
of group was unreliable, F < 1.
BMI
The mean BMI was lower in participants who
were low in body dissatisfaction as compared
to participants who were high in body dissat-
isfaction, 20.17 vs. 23.14, respectively, t(62) =
4.22, p < .001, d = 1.05.
Correlational Analyses
Table 1 provides an overview of the descrip-
tive statistics of and pairwise correlations
actual and ideal body image, explicit body
Heider et al: Body Dissatisfaction Revisited
dissatisfaction as measured by the EDI, the
ratings of the RRT statements, and the BMI.
Due to our sampling method (i.e., two groups
with either a high or a low level of body dis-
satisfaction), most variables were not nor-
mally distributed. Accordingly, Spearman’s
rank order correlations were computed. The
actual-RRT correlated significantly with the
explicit measures of actual body image (i.e.,
actual-CDRS scores, actual-RRT statement
ratings, and BMI scores), but not with the
explicit measures of ideal body image (i.e.,
ideal-CDRS scores and ideal-RRT statement
ratings). The ideal-RRT, in contrast, correlated
significantly with Ideal-CDRS scores, but was
unrelated to explicit measures of actual body
image (i.e., actual-CDRS scores, actual-RRT
statement ratings, and BMI scores). Finally,
it was observed that the actual-RRT and
the ideal-RRT did not correlate with one
another.
Hierarchical Logistic Regression Analyses
To further validate the hypothesis that the
two RRTs captured different implicit beliefs,
a logistic regression was performed in which
the two RRT scores as well as their interaction
were used as predictors of whether a partici-
pant had been assigned to the group with a
high or low degree of body dissatisfaction.
On the basis of the full model, 67.2 % of all
participants were classified correctly, χ2(3) =
12.98, p < .001, Nagelkerke R² = .25. Whereas
the actual-RRT contributed significantly
to the prediction of group membership,
Wald = 7.32, p < .05, OR = 0.42, the effects for
the ideal-RRT and the interaction term were
marginally significant only, Wald = 2.83,
p = .09, OR = 1.69, and Wald = 2.97, p = .09,
OR = 0.58, respectively.
Subsequently, we examined the predic-
tive validity of the RRT data over and above
the predictive validity of the (mean) explicit
ratings of the statements used in the two
RRTs. In a first step, the rating scores for the
actual-RRT, the rating scores for the ideal-
RRT, as well as the interaction between the
two ratings scores were used as predictors of
167
group membership. Results showed that the
rating scores for the actual-RRT were highly
predictive of group membership, Wald =
5.24, p = .05, OR = 0.011, whereas the other
predictors were not (i.e., Wald < 1.80, ps > .18).
Accordingly, only the rating scores for the
actual-RRT were retained for a hierarchical
analysis in which the two RRT scores and
their interaction were entered in a second
step. While the overall model fit was very
high, χ2(4) = 67.75, p < .001, Nagelkerke
R² = .87 (93.8 % correct classifications),
the added value of the RRT measures was
­negligible, χ2(3) = 2.75, p > .40.
A similar analysis was performed for the
CDRS measures. Both the actual-CDRS scores
and the ideal-CDRS scores were good predic-
tors of group membership, Wald = 16.68,
p < .001, OR = 39.69, and Wald = 8.82,
p < .005, OR = 0.12. In contrast, the inter-
action between the two CDRS measures
was unrelated to group membership, Wald
< 1.76, p > .18, and was thus dropped from the
model. When adding the two RRT ­measures
and their interaction in a second step, model
fit increased to a significant extent, χ2(3) =
8.48, p < .05. Final results showed, however,
that only the actual-RRT score tended to be
predictive of group membership over and
above to the two CDRS measures, Wald =
3.80, p = .051, OR = 0.36. On the basis of the
full model, 90.6 % of all participants were
classified correctly, χ2(5) = 55.17, p < .001,
Nagelkerke R² = .77.
Finally, we examined the predictive validity
of the RRT measures over and above the BMI.
While the BMI alone was a reliable predictor
of group membership, Wald = 10.72, p <
.05, OR = 3.83, the inclusion of the two RRT
measures and their interaction resulted in a
significantly better model fit, χ2(3) = 10.35,
p < .05. Whereas the ideal-RRT contributed
significantly to the prediction of group mem-
bership, Wald = 4.44, p < .05, OR = 2.34, the
effects for the actual-RRT and the interaction
term just missed significance, Wald = 2.79,
p = .10, OR = 0.55, and Wald = 3.71, p = .05,
OR = 0.48, respectively. On the basis of the
168
full model, 76.6 % of all participants were
classified correctly, χ2(4) = 27.17, p < .001,
Nagelkerke R² = .46.
Discussion
The degree to which people are dissatis-
fied with their own body is assumed to
reflect a (perceived) discrepancy between
the actual and the ideal body image (e.g.,
Cooper & Taylor, 1988; Strauman et al., 1991;
Williamson et al., 1993). Consistent with
this idea, Heider et al. (2015) observed that
implicit beliefs about actual and ideal body
image, as measured by the IRAP, were differ-
ent in participants who reported either a low
or a high degree of body dissatisfaction. In
the present study, using the RRT (De Houwer
et al., 2015) as a measure to capture implicit
beliefs, we replicated the findings by Heider
et al. (2015). Specifically, we observed that
the implicit belief to be thin was more pro-
nounced for participants low in body dissat-
isfaction as compared to participants high
in body dissatisfaction. The implicit desire
to be thin, in contrast, tended to be more
pronounced in participants high in body dis-
satisfaction as compared to participants low
in body dissatisfaction. These findings dem-
onstrate that the RRT, like the IRAP, is capa-
ble of distinguishing between closely related
beliefs that differ only in their relational
component (i.e., ‘I am thin’ vs. ‘I want to be
thin’). Additional findings corroborated this
inference. The actual-RRT correlated with an
explicit measure of actual body image but
not with an explicit measure of ideal body
image. Conversely, the ideal-RRT correlated
with an explicit measure of ideal body image
but not with an explicit measure of actual
body image. Taken together with the obser-
vation that both RRTs were unrelated to each
other, these findings strongly support the
conclusion that both RRT measures, despite
their structural similarity, captured different
beliefs.
The ability of the RRT to pick up inter-
individual differences in implicit beliefs is
an important finding as complex relational
Heider et al: Body Dissatisfaction Revisited
knowledge is rather difficult to capture using
more traditional implicit measures such as
the IAT (Greenwald et al., 1998). In the con-
text of body dissatisfaction, for example,
one might construct an IAT in which words
referring to the self or well-known other
persons are either mapped on the same
response key or a different response key as
words referring to thinness and overweight
(i.e., Bluemke & Friese, 2012). While such
an approach would allow for an assessment
of the degree to which the self is linked in
memory to the concept of thinness/thick-
ness, it would remain unknown whether this
link reflects and actual state (i.e., ‘I am thin’)
or a desired state (i.e., ‘I want to be thin’). If
such a discrimination is the objective, rela-
tional implicit measures are needed. The pre-
sent data thus show that the RRT can be a
valuable addition to the traditional toolbox
of implicit measures, especially in the field of
body dissatisfaction.
The present results also go beyond the
findings of Heider et al. (2015), in sev-
eral ways. First, while Heider et al. (2015)
observed no incremental predictive validity
of the IRAP over and above explicit meas-
ures of body dissatisfaction, we observed
that the predictive validity RRT measures
were predictive of body dissatisfaction
over and above the CDRS and the BMI.
This observation is important as it suggests
that the RRT measures developed here may
eventually be used to predict important
behavioral outcomes (e.g., relapse in eat-
ing disorders). Still, the observation that
there was no evidence of incremental pre-
dictive validity over and above the explicit
ratings is somewhat problematic from this
perspective. It should be noted, however,
that implicit measures are expected to
be particularly useful in situations where
explicit measures are biased as the result of
social desirability or self-presentation con-
cerns. It could thus be hypothesized that
the added value of the RRT scores over and
above the explicit ratings might surface in
participants who are motivated to respond
Heider et al: Body Dissatisfaction Revisited 169

in a biased manner. Second, while Heider
et al. (2015) simply observed a correlation
between body dissatisfaction and the scores
of the ideal-IRAP, the present data suggest
that the degree to which the implicit desire
to be thin is predictive of body dissatisfac-
tion might be dependent on the extent to
which one beliefs to be thin at the implicit
level. As can be seen in Figure 1, for partici-
pants whose actual-RRT score was indicative
of the belief to be overweight, body dissat-
isfaction increased as the implicit desire to
be thin increased. In contrast, participants
whose actual-RRT scores were indicative of
the belief to be thin, the implicit desire to be
thin was unrelated to group membership.
This data pattern is in perfect accordance
with the conceptualization of body dis-
satisfaction as the (self-perceived) discrep-
ancy between actual and ideal body image.
According to such a viewpoint, the desire to
be thin should promote body dissatisfaction
only if a person does not possess the belief
of being thin. It must be noted, however,
that the interaction between implicit actual
and ideal body image just missed signifi-
cance, so some caution is advised in ­drawing
these conclusions.

Figure 1: Probability of group membership (high vs. low degree of body dissatisfaction) as
a function of the implicit desire to possess a thin body (i.e., D1 scores for the ideal-RRT),
­separated for low (i.e., 1 SD below average), average, and high (i.e., 1 SD above average)
degrees of the implicit belief that one is thin (i.e., D1 scores for the actual-RRT).
170
The present study already has important
practical implications. Albeit the internal
consistency estimates for the actual-RRT and
the ideal-RRT were relatively modest (Rsb =
0.49 and Rsb = 0.57, respectively), they were
clearly much higher than those reported by
Heider et al. (2015) for the actual-IRAP and
the ideal-IRAP (Rsb = 0.32 and Rsb = 0.24,
respectively). In addition, in both RRTs, the
mean response latency was well below 2000
ms and the overall error rate was smaller
than 15 %. Finally, less than 6 % of the par-
ticipants were outliers in terms of their mean
speed of responding or error rate and the
overall results were unaffected by the inclu-
sion or exclusion of these participants. Taken
together, these observations add further
weight to the idea that, in comparison to the
IRAP, the RRT is an easy-to-complete instru-
ment that might be useful as a diagnostic
instrument also outside of the laboratory.
It may be noted, however, that more
research would be needed to substantiate the
causal nature of the relationship between, on
the one hand, implicit beliefs about actual
and ideal body image and, on the other hand,
important behavioral outcomes such as the
occurrence or maintenance of eating disor-
ders. As an experimental approach rather
than a correlational approach is needed to
address this issue, it seems particularly inter-
esting to start examining how and to what
extent (complex) implicit beliefs can be
changed. Not only would such an approach
advance our general understanding of the
relationship between implicit beliefs and
behavioral outcomes, it is also likely to result
in new therapeutic intervention strategies.
Additional Files
The additional files for this article can be
found as follows:
• Appendix. [Description]. DOI: https://doi.
org/10.5334/pb.362.s1
Funding Information
Preparation of this paper was supported by
Methusalem Grant BOF16/MET_V/002 of
Ghent University awarded to Jan De Houwer.
Heider et al: Body Dissatisfaction Revisited
Competing Interests
The reported research was in part conducted
while Adriaan Spruyt was a Postdoctoral
Fellow of the Flemish Research Foundation
(FWO – Vlaanderen).
References
Allebeck, P., Hallberg, D., & Espmark, S.
(1976). Body image: An apparatus for
measuring disturbances in estimation of
size and shape. Journal of Psychosomatic
Research, 20, 583–589. DOI: https://doi.
org/10.1016/0022-3999(76)90060-X
American Psychiatric Association. (2013).
Diagnostic and statistical manual of
mental disorders (5th ed.). Arlington, VA:
American Psychiatric Publishing. DOI:
https://doi.org/10.1176/appi.books.978
0890425596
Barnes-Holmes, D., Barnes-Holmes, Y.,
Power, P., Hayden, E., Milne, R., &
Stewart, I. (2006). Do you really know
what you believe? Developing the Implicit
Relational Assessment Procedure (IRAP)
as a direct measure of implicit beliefs. The
Irish Psychologist, 32, 169–177.
Barnes-Holmes, D., Murtagh, L., Barnes-
Holmes, Y., & Stewart, I. (2010). Using
the Implicit Association Test and the
Implicit Relational Assessment Procedure
to measure attitudes toward meat and
vegetables in vegetarians and meat-eaters.
The Psychological Record, 60, 287–305.
DOI: https://doi.org/10.1007/BF0339
5708
Bessenoff, G. R., & Sherman, J. W. (2000).
Automatic and controlled components
of prejudice toward fat people: Evalua-
tion versus stereotype activation. Social
Cognition, 18, 329–353. DOI: https://doi.
org/10.1521/soco.2000.18.4.329
Bluemke, M., & Friese, M. (2012). On the
validity of idiographic and generic self-
concept implicit association tests: A core-
concept model. European Journal of Per-
sonality, 26, 515–528. DOI: https://doi.
org/10.1002/per.850
Campbell, C., Barnes-Holmes, Y., Barnes-
Holmes, D., & Stewart, I. (2011). Exploring
screen presentations in the implicit
Heider et al: Body Dissatisfaction Revisited
relational assessment procedure (IRAP).
International Journal of Psychology and
Psychological Therapy, 11, 377–388.
Cash, T. F. (1990). The psychology of physical
appearance: Aesthetics, attributes, and
images. In Cash, T. F. & Szymanski, M. L.
(Eds.), Body images: Development, devi-
ance, and change (pp. 51–79). New York:
Guilford Press.
Clausen, L., Rosenvinge, J. H., Friborg, O., &
Rokkedal, K. (2011). Validating the
eating disorder inventory-3 (EDI-3): A
comparison between 561 female eating
disorders patients and 878 females from
the general population. Journal of Psycho-
pathology and Behavioral Assessment, 33,
101–110. DOI: https://doi.org/10.1007/
s10862-010-9207-4
Cooper, P. J., & Taylor, M. J. (1988). Body-
image disturbance in bulimia-nervosa.
British Journal of Psychiatry, 153, 32–36.
Cronbach, L. (1990). Essentials of psychologi-
cal testing. New York: Harper & Row.
Degner, J., & Wentura, D. (2009). Not
everybody likes the thin and despises the
fat: One’s weight matters in the auto-
matic activation of weight-related social
evaluations. Social Cognition, 27, 202–221.
DOI: https://doi.org/10.1521/soco.2009.27.
2.202
De Houwer, J. (2003). The extrinsic affec-
tive Simon task. Experimental Psychology,
50, 77–85. DOI: https://doi.org/10.1027/
1618-3169.50.2.77
De Houwer, J., Heider, N., Spruyt, A.,
Roets, A., & Hughes, S. (2015). The rela-
tional responding task: toward a new
implicit measure of beliefs. Frontiers in
Psychology, 6, ARTN 319. DOI: https://
doi.org/10.3389/fpsyg.2015.00319
De Houwer, J., Teige-Mocigemba, S.,
Spruyt, A., & Moors, A. (2009). Implicit
measures: A normative analysis and review.
Psychological Bulletin, 135, 347–368. DOI:
https://doi.org/10.1037/a0014211
Drake, C. E., Kellum, K. K., Wilson, K. G.,
Luoma, J. B., Weinstein, J. H., &
Adams, C. H. (2010). Examining the
171
implicit relational assessment procedure:
Four preliminary studies. The Psychological
Record, 60, 81–100. DOI: https://doi.
org/10.1007/BF03395695
Fairburn, C. G., & Harrison, P. J. (2003).
Eating disorders. The Lancet, 361,
407–416. DOI: https://doi.org/10.1016/
S0140-6736(03)12378-1
Fazio, R. H., Jackson, J. R., Dunton, B. C., &
Williams, C. J. (1995). Variability in
automatic activation as an unobtrusive
measure of racial-attitudes: A bona-fide
pipeline? Journal of Personality and Social
Psychology, 69, 1013–1027. DOI: https://
doi.org/10.1037//0022-3514.69.6.1013
Freeman, R. J., Thomas, C. D., Solyom, L., &
Hunter, M. A. (1984). A modified video cam-
era for measuring body-image distortion:
Technical description and reliability. Psycho-
logical Medicine, 14, 411–416. DOI: https://
doi.org/10.1017/S0033291700003652
Garner, D. M., Olmstead, M. P., & Polivy, J.
(1983). Development and validation of a
multidimensional eating disorder inventory
for anorexia-nervosa and bulimia. Interna-
tional Journal of Eating Disorders, 2, 15–34.
DOI: https://doi.org/10.1002/1098-108X
( 19 8 3 21 ) 2 : 2 < 15 : : A I D - E AT 2 2 6 0 0
20203>3.0.CO;2-6
Greenwald, A. G., & Banaji, M. R. (1995).
Implicit social cognition: Attitudes, self-
esteem, and stereotypes. Psychological
Review, 102, 4–27. DOI: https://doi.org/
10.1037/0033-295X.102.1.4
Greenwald, A. G., McGhee, D. E., &
Schwartz, J. L. K. (1998). Measuring
individual differences in implicit c­ ognition:
The implicit association test. ­ Journal of
Personality and Social Psychology, 74,
1464–1480. DOI: https://doi.org/10.1037/
0022-3514.74.6.1464
Greenwald, A. G., Nosek, B. A., & Banaji, M. R.
(2003). Understanding and using the
implicit association test: I. An improved
scoring algorithm. Journal of Personality
and Social Psychology, 85, 197–216.
DOI:  https://doi.org/10.1037/0022-
3514.85.2.197
172
Heider, N., Spruyt, A., & De Houwer, J.
(2015). Implicit beliefs about ideal body
image predict body image dissatisfaction.
Frontiers in Psychology, 6, ARTN 1402.
DOI: https://doi.org/10.3389/fpsyg.2015.
01402
Holtgraves, T. (2004). Social desirability
and self-reports: Testing models of
socially desirable responding. Personality
and Social Psychology Bulletin, 30, 161–172.
DOI: https://doi.org/10.1177/0146167203
259930
Hughes, S., & Barnes-Holmes, D. (2013).
A functional approach to the study of
implicit cognition: The IRAP and the REC
model. In Dymond, S. & Roche, B. (Eds.),
Advances in Relational Frame Theory &
Contextual Behavioural Science: Research &
Applications (pp. 97–126). Oakland, CA:
New Harbinger Publications.
Hughes, S., Barnes-Holmes, D., & Vahey, N.
(2012). Holding on to our functional roots
when exploring new intellectual islands:
A voyage through implicit cognition
research. Journal of Contextual Behavio-
ral Science, 1, 17–38. DOI: https://doi.
org/10.1016/j.jcbs.2012.09.003
Juarascio, A. S., Forman, E. M., Timko, C. A.,
Herbert, J. D., Butryn, M., & Lowe, M.
(2011). Implicit internalization of the
thin ideal as a predictor of increases
in weight, body dissatisfaction, and
disordered eating. Eating Behaviors, 12,
207–213. DOI: https://doi.org/10.1016/j.
eatbeh.2011.04.004
Parling, T., Cernvall, M., Stewart, I., Barnes-
Holmes, D., & Ghaderi, A. (2012). Using
the implicit relational assessment proce-
dure to compare implicit pro-thin/anti-
fat attitudes of patients with anorexia
nervosa and non-clinical controls. Eating
Disorders, 20, 127–143. DOI: https://doi.
org/10.1080/10640266.2012.654056
Payne, B. K., Cheng, C. M., Govorun, O., &
Stewart, B. D. (2005). An inkblot for
attitudes: Affect misattribution as implicit
measurement. Journal of Personality and
Social Psychology, 89, 277–293. DOI:
Heider et al: Body Dissatisfaction Revisited
https://doi.org/10.1037/0022-3514.89.
3.277
Ratcliff, R. (1993). Methods for dealing
with reaction-time outliers. Psychological
­Bulletin, 114, 510–532. DOI: https://doi.
org/10.1037/0033-2909.114.3.510
Remue, J., De Houwer, J., Barnes-­
Holmes, D., Vanderhasselt, M. A., &
De Raedt, R. (2013). Self-esteem revisited:
Performance on the implicit relational
assessment procedure as a measure of
self-versus ideal self-related cognitions
in dysphoria. Cognition and Emotion, 27,
1441–1449. DOI: https://doi.org/10.108
0/02699931.2013.786681
Remue, J., Hughes, S., De Houwer, J., &
De Raedt, R. (2014). To be or want to
be: Disentangling the role of actual
versus ideal self in implicit self-esteem.
Plos One, 9, ARTN e108837. DOI: https://
doi.org/10.1371/journal.pone.0108837
Roddy, S., Stewart, I., & Barnes-Holmes, D.
(2010). Anti-fat, pro-slim, or both? Using
two reaction-time based measures to
assess implicit attitudes to the slim and
overweight. Journal of Health Psychology,
15, 416–425. DOI: https://doi.org/10.1177/
1359105309350232
Roddy, S., Stewart, I., & Barnes-Holmes, D.
(2011). Facial reactions reveal that slim
is good but fat is not bad: Implicit and
explicit measures of body-size bias.
­European Journal of Social Psychology, 41,
688–694. DOI: https://doi.org/10.1002/
ejsp.839
Slade, P. D., & Russell, G. F. M. (1973). Aware-
ness of body dimensions in anorexia
nervosa cross-sectional and longitudi-
nal studies. Psychological Medicine, 3,
188–199. DOI: https://doi.org/10.1017/
S0033291700048510
Spruyt, A., Clarysse, J., Vansteenwegen, D.,
Baeyens, F., & Hermans, D. (2010).
Affect 4.0 a free software package for
implementing psychological and psycho-
physiological experiments. Experimental
Psychology, 57, 36–45. DOI: https://doi.
org/10.1027/1618-3169/a000005
Heider et al: Body Dissatisfaction Revisited 173

Stice, E. (2001). A prospective test of the
dual-pathway model of bulimic pathol-
ogy: Mediating effects of dieting and
negative affect. Journal of Abnormal Psy-
chology, 110, 124–135. DOI: https://doi.
org/10.1037/0021-843X.110.1.124
Stice, E. (2002). Risk and maintenance fac-
tors for eating pathology: A meta-­analytic
review. Psychological Bulletin, 128,
825–848. DOI: https://doi.org/10.1037/
0033-2909.128.5.825
Strauman, T. J., Vookles, J., Berenstein, V.,
Chaiken, S., & Higgins, E. T. (1991). Self-
Discrepancies and vulnerability to body
dissatisfaction and disordered e­ating.
Journal of Personality and Social Psy-
chology, 61, 946–956. DOI: https://doi.
org/10.1037/0022-3514.61.6.946
Thompson, J. K., & Stice, E. (2001). Thin-ideal
internalization: Mounting evidence for a
new risk factor for body-image disturbance
and eating pathology. Current Directions
in Psychological Science, 10, 181–183. DOI:
https://doi.org/10.1111/1467-8721.00144
Thompson, M. A., & Gray, J. J. (1995). Develop-
ment and validation of a new body-image
assessment scale. Journal of Personality
Assessment, 64, 258–269. DOI: https://
doi.org/10.1207/s15327752jpa6402_6
Williamson, D. A., Gleaves, D. H.,
Watkins, P. C., & Schlundt, D. G. (1993).
Validation of self ideal body-size discrep-
ancy as a measure of body dissatisfaction.
Journal of Psychopathology and Behavio-
ral Assessment, 15, 57–68. DOI: https://
doi.org/10.1007/BF00964324

How to cite this article: Heider, N., Spruyt, A., and De Houwer, J. (2018). Body Dissatisfaction
Revisited: On the Importance of Implicit Beliefs about Actual and Ideal Body Image. Psychologica
Belgica, 57(4), 158–173, DOI: https://doi.org/10.5334/pb.362

Submitted: 09 September 2016 Accepted: 02 March 2017 Published: 04 January 2018

Copyright: © 2018 The Author(s). This is an open-access article distributed under the terms of the
Creative Commons Attribution 4.0 International License (CC-BY 4.0), which permits unrestricted use,
distribution, and reproduction in any medium, provided the original author and source are credited.
See http://creativecommons.org/licenses/by/4.0/.

Psychologica Belgica is a peer-reviewed open access journal

OPEN ACCESS
published by Ubiquity Press.